1. Field of the Invention
The present invention relates to a fabrication process of solar cell. More specifically, the present invention relates to a process of fabricating a solar cell suitably applicable to a solar cell comprising stacked thin films of polycrystalline silicon on a cheap substrate.
2. Related Background Art
Solar cells are widely studied and some are in practical use, as driving energy sources of various devices or as power supplies connected to the commercial power supply system.
A requirement for the solar cells in respect of cost is that an element can be formed on a cheap substrate like metal. On the other hand, silicon is normally used as a semiconductor for making the solar cells. Among others, single-crystal silicon is most excellent from the viewpoint of efficiency for converting light energy to electromotive force, i.e., from the viewpoint of photoelectric conversion efficiency. It is, however, said that amorphous silicon is more advantageous from the viewpoints of increase in area and decrease in cost. In recent years, use of polycrystalline silicon is under study, for the purpose of achieving low cost comparable to that of amorphous silicon and high energy conversion efficiency comparable to that of single-crystal silicon. In conventionally proposed methods as to such single-crystal silicon and polycrystalline silicon, a plate-shaped substrate was obtained by slicing a massive crystal. It was thus not easy to decrease the thickness of the substrate to below 0.3 mm. Therefore, the substrate had the thickness more than necessary for sufficient absorption of light quantity, and effective utilization of material was not enough. Namely, the substrate needed to be thinned more in order to further decrease the cost.
Proposed as a production method of polycrystalline silicon substrate with the aim of decreasing the cost was a method for forming a silicon sheet by a spin method of pouring a liquid droplet of molten Si into a mold. This method achieved the minimum thickness of about 0.1 to 0.2 mm, but the decrease of thickness was not sufficient as compared with the thickness of film (20 to 50 μm) necessary and sufficient for absorption of light as crystalline Si. In addition, this thinning method had the problem that it became difficult for the silicon sheet itself to maintain the strength as a substrate, so that another cheap substrate was inevitably necessitated for supporting the silicon sheet.
A report was made about attempts to make the solar cell by forming a substrate of metal-grade silicon and thereafter forming a silicon layer having a thickness necessary and sufficient for absorption of light thereon by the liquid-phase growth method (T. F. Ciszek, T. H. Wang, X. Wu, R. W. Burrows, J. Alleman, C. R. Schwerdtfeger and T. Bekkedahl, “Si thin layer growth from metal solution on single-crystal and cast metallurgical-grade multicrystalline Si substrates,” 23rd IEEE Photovoltaic specialists Conference, (1993) p. 65).
In the above-stated method wherein the silicon layer was made on the low-purity silicon substrate of metal-grade silicon, using copper, aluminum, or tin as a metal solvent, however, the metal as a solvent was left mainly in grain boundaries in either case, because etchback for removing a native oxide film was carried out in the initial stage of growth. Therefore, characteristics of the solar cell were not sufficient. For the purpose of solving this problem, another method was reported which used a copper and aluminum alloy as the solvent without carrying out the etchback (T. H. Wang, T. F. Ciszek, C. R. Schwerdtfeger, H. Moutinho, R. Matson., “Growth of silicon thin layers on cast MG-Si from metal solutions for solar cells,” Solar Energy Materials and Solar Cells 41/42 (1996), p. 19), but this method has problems including complex composition control of the alloy in regard to mass production.
A method for making the substrate of metal-grade silicon is also the same as the conventional polycrystalline process for forming an ingot by the cast method and slicing it to obtain the plate-shaped substrate. Such a method fails to make use of the merit of metal-grade silicon as a cheap material.